Abstract Book

S156

ESTRO 37

contrast, the side-by-side scoring showed a difference in plan quality of +18.1 ± 41, pointing at an enhanced quality for VMAT auto , which was statistically significant (p<0.001) and in line with DVH comparisons between VMAT auto and VMAT clin . The large difference in standard deviations (above: 12 vs. 41) shows that in the side-by- side comparisons, the clinicians were much more convinced of plan quality differences. Figure 1 shows the overall poor correlation between separate and side-by- side plan scoring (R 2 =0.22), with clinician-specific R 2 - values of 0.02, 0.12, 0.27, 0.33 and 0.75.

the MR-Linac. It shows that the challenges arising from the presence of the MR housing in the set-up are not an impediment for the implementation of EPID dosimetry in the MR-Linac. The adaptation of the model for all gantry angles, and the reconstruction of the dose distribution in the attenuated area of the beam are ongoing work.

Proffered Papers: PH 6: Treatment planning 1

OC-0299 Inconsistencies in clinicians‘ final treatment plan evaluations – a need for automation support B. Heijmen 1 , P. Bonomo 2 , G. Goldner 3 , A. Henry 4 , F. Lohr 5 , G. Simontacchi 2 , P. Voet 6 , D. Fransen 1 , J. Penninkhof 1 , M. Milder 1 , A. Akhiat 6 , M. Casati 2 , D. Georg 3 , J. Lilley 4 , L. Marrazzo 2 , S. Pallotta 2 , R. Pellegrini 7 , Y. Seppenwoolde 3 , V. Steil 8 , F. Stieler 8 , S. Wilson 4 , S. Breedveld 1 1 Erasmus MC Cancer Institute, Radiation Oncology, Rotterdam, The Netherlands 2 Azienda Ospedaliero-Universitaria Careggi, Radiation Oncology, Florence, Italy 3 Medical University of Vienna- Christian Doppler Laboratory for Medical Radiation Physics for Radiation Oncology / AKH Wien, Radiation Oncology, Vienna, Austria 4 Leeds Cancer Centre- St James's University Hospital, Radiation Oncology, Leeds, United Kingdom 5 Azienda Ospedaliero-Universitaria Modena- Policlinico, UO Radioterapia- Dip. di Oncologia, Modena, Italy 6 Elekta AB, Elekta, Rotterdam, The Netherlands 7 Elekta AB, Elekta, Milano, Italy 8 University Medical Center Mannheim- Heidelberg University, Radiation Oncology, Mannheim, Germany Purpose or Objective In current interactive trial-and-error treatment planning, final plan quality depends on the skills of the planner and on allotted time for planning. At the end of the planning process, the treating clinician needs to give final plan approval to start with treatment. Ideally, suboptimal treatment plans are recognized by the clinician to prevent treatment with them. In this study, we systematically investigated the consistency of clinicians‘ plan evaluations by comparison of: A. independent plan quality scoring by a clinician of alternative plans of patients, separated in time, with B. side-by-side, direct comparison of alternative plans. Material and Methods Five clinicians from 4 treatment centers in 4 European countries participated in the study. In each center, treatment plans of 20 prostate cancer patients were evaluated (total 80 patients). For each patient, 2 VMAT plans were available, the clinically delivered plan (VMAT clin ) and an alternative plan that was automatically generated with a system for automated planning (VMAT auto ). In each center, plan quality was scored in two sessions: independent, subsequent quality scoring for each of the 20x2 plans in a random order (A.), and, for each of the 20 patients, side-by-side plan comparison and direct scoring of the quality difference between the 2 VMAT plans (B.). Both for A. and B., scoring was performed using visual analogue scales. We compared for each patient the plan quality difference, calculated as the difference in the independent scores for VMAT clin and VMAT auto obtained with approach A., with the plan quality difference obtained by direct plan comparison. For both approaches, plan quality difference-scores ranged from - 100 to +100. Results In the independent plan scoring, the average plan quality difference was -2.5 ± 12, i.e. a small advantage for VMAT clin , which was borderline significant, p=0.06. In

Conclusion Large inconsistencies were observed in clinicians‘ plan evaluations. Such evaluations may often not prevent treatment with a suboptimal plan. Automated planning and/or automated plan QA are needed to better guarantee high quality treatment. OC-0300 Linac MRI guided SBRT treatment in pancreatic cancer: dosimetric evaluation of a new technology D. Cusumano 1 , S. Menna 1 , L. Boldrini 2 , S. Teodoli 1 , E. Placidi 1 , G. Chiloiro 2 , L. Placidi 1 , F. Greco 1 , G. Stimato 1 , F. Cellini 2 , V. Valentini 2 , M. De Spirito 1 , L. Azario 1 1 Fondazione Policlinico Universitario A.Gemelli, Unità Operativa Complessa di Fisica Sanitaria, Roma, Italy 2 Fondazione Policlinico Universitario A.Gemelli, Area di Radioterapia Oncologica, Roma, Italy Purpose or Objective Hypofractionated RT today plays a relevant role in different moments of the clinical management of pancreatic cancer patients. MR guided adaptive RT (MR- ART) is considered a new promising resource in this context, as it allows to online modify the dose distribution according to therapy volumes position of the day. Aim of this in silico study is to evaluate the dosimetric performance of the first 6MV MR-Linac authorised for clinical treatments (MRIdian Linac, ViewRay) in case of SBRT pancreas treatments. The results have been compared with those obtained with a 6MV Linac specifically designed for SBRT treatments (TrueBeam Edge, Varian). Material and Methods Eight patients affected by locally advanced pancreatic cancer were considered.